Validation through specialised mechanical testing - the Abington solution
After obtaining his engineering degree in Germany and his Doctorate at the Swiss Institute of Technology, EurIng Dr Christoph Wiesner joined TWI in 1991. Following seven years as a Project Leader, Section Manager and Deputy Head, he was appointed Head of TWI's Structural Integrity Department in 1998. He collaborates on several BSI and CEN Technical Committees and is author of more than 60 publications.
The importance of fatigue and fracture as potential sources of failure in welded structures and components is reflected in the high level of consultancy and research activity at TWI over the last 50 years. Christoph Wiesner spotlights some of the Structural Integrity Department's recent specialised testing operations.
In recent times, research into fracture mechanics has developed to encompass new materials and concepts. The predictive capability of computer modelling of structural behaviour has improved markedly such that it may be considered as a future alternative for routine testing. However, for new joining processes, novel design features and state-of-the-art flaw assessment procedure developments to be accepted with confidence, validation testing remains a crucial part of the qualification process.
This article presents a series of recent examples where TWI's unique large-scale component test facilities were used to develop and carry out mechanical tests for validation purposes.
Examples of specialised testing in Sid's laboratory
Validation tests for friction stir welding process
The invention of the friction stir welding (FSW) process at TWI has excited the aerospace industry around the world with the prospect of high integrity welds in aluminium alloys that were formerly considered unweldable. Recent work has been concerned with the fracture resistance of friction stir welds for rocket components at cryogenic temperatures.
Figure 1 shows a wide plate test in a newly commissioned tension test rig. This has a load capacity of greater than 100 tonnes. A special feature of the new test rig is that it enables wide plate specimens to be tested in a horizontal plane, which is a great advantage for tests at low temperatures. For example, one of the tests, a 700mm wide by 6mm thick middle crack tension specimen was immersed in a shallow bath of liquid nitrogen for a low temperature test. The rig allowed testing of the first friction stir welded Al alloy wide plate at precisely -196°C.
Proof testing of aircraft landing gear
TWI carried out a series of tests on large welded undercarriage components. A machine of 180 tonnes capacity with daylight between crossheads of approximately 1.5m ( Fig.2) was available, and testing began within two days of the first contact.
The components tested formed part of the landing gear for the McDonnell Douglas MD11 aircraft. Fabricated in ultra-high-strength steel, they were examined by magnetic particle inspection (MPI) and heat treated before transfer to TWI for proof loading. Specified as being to 60% of UTS, this test on two designs of component required application of tensile loads of 140 to 160 tonnes to parts over one metre long.
Fatigue design of mechanical fasteners
Although most of TWI's fatigue projects involve welded joints, expertise in other methods of joining is increasing. For example, there is strong activity in the field of adhesive bonding, including fatigue and fracture studies, and projects directed at the structural integrity of automotive products have included comparative fatigue tests on welded, bonded and mechanically fastened joints between steel or aluminium sheets.
A feature in all cases is that fatigue testing is most commonly performed on simple lap joint specimens. This is also the case with bolted joints. A recent Eureka project included such fatigue tests on friction grip-bolted joints in aluminium alloys. Subsequently, this led to a study of the design of such joints in real structures based on fatigue data obtained from the simple lap specimens. The study included finite element analysis and fatigue testing of structural components ( Fig.3) under both bending and torsion loading. The work highlighted the problem of relating the fatigue behaviour of lap specimens to real structures, indicating that similar work is needed for the other types of joint.
High rate fracture toughness testing
To identify candidate tube materials and welds for the design of storage tubes for spent nuclear fuel elements ( Fig.4), an experimental programme to determine the static and dynamic tensile and fracture toughness properties was carried out. TWI's 10m drop-weight tower was used for the high loading rate tests, which allowed an impact speed of up to 30mph ( Fig.5).
Fracture toughness test results on parent steels and welds revealed that an increase in loading rate causes a shift in the fracture toughness transition curve. For one steel this resulted in a change in fracture mode from fully ductile to brittle.
These test results demonstrated the importance of assessing design scenarios under realistic conditions. Fracture toughness properties are likely to be markedly less forgiving to the existence of weld flaws at high loading rates.
Toughness requirements for steel bridges
TWI has carried out a programme of work to devise rules for avoidance of brittle fracture in both new and existing steel bridges. Design and assessment rules were formulated from first principles using a fracture mechanics route, and then calibrated against a series of large-scale laboratory tests and actual failure case studies.
Large-scale test specimens were made from 50mm thick steel plates with various welded attachments to represent different types of stress concentration ( Fig.6). A cover plate, a transverse stiffener and an edge attachment were all fabricated, these being typical of details used in bridge design. Artificial flaws were introduced in the area of the attachment by machining and fatigue pre-cracking, and low-temperature fracture tests were then carried out using TWI's 4000 tonne wide plate test rig. Results were analysed in accordance with the proposed bridge design rules and with the UK defect acceptance procedure BS 7910.
Full-scale validation tests on lighting columns
Lighting columns up to 18 metres high have been the subject of study at TWI following exceptional storms which caused the collapse of a dozen lighting columns. Ninety more had to be cut down when it was revealed that they showed signs of storm damage. The winds to which they had been subjected were only about half the UK design speed for such columns. There was a need for the relevant design and fabrication standard to be improved to prevent these failures which were due to fatigue.
Thirty-six sample lighting columns were randomly selected from various makers for investigation at Abington. The full-scale components were fatigue tested in TWI's fatigue laboratory and several failure vessels and locations were identified (see Fig.7). As a direct result of this work, British Standards rules are now modified to be consistent with the results obtained.
Testing of escalator steps
Escalators, in common with most other types of public service equipment, have to be thoroughly designed, tested and manufactured to minimise the risk of mechanical failure. The consequences of failure in an escalator system could be very serious indeed, and there have been instances where passengers have been injured as a result of a failure in the system.
The relevant standard specifies static and fatigue tests for escalator steps, and in recent years TWI has carried out these tests and many others on different designs of step for a number of manufacturers.
Static testing involves applying a single downward load on the top of the step. Elastic deflections under load must be within certain limits, and very little permanent distortion is permitted. Fatigue testing involves applying a similar load for a minimum of five million cycles, after which the step must show no sign of fatigue cracking or other damage.
The test loads are severe, the maximum applied force being 0.3 tonnes. Control of deflection under static load ensures that nothing can jam or seize up, and if a step survives the fatigue test, one may be confident that in service the step will have an infinite fatigue life in service (see Fig.8).
TWI's test facilities are ideal for performing this type of testing and, since one design of escalator step is invariably slightly different from another, a collection of test frames is available.
Burst testing of pipeline containing simulated corrosion defects
TWI was asked to investigate the failure pressure of an API 5L X52 grade pipe subjected to severe corrosion pitting. Using finite element analysis (FEA), the depth of the pitting was shown to have comparatively little influence on the burst pressure for corrosion flaws seen in practice. A specimen of pipe with a representative defect machined on the inside surface was burst tested to provide additional assurance of fitness-for-service and support the finite element test results. The actual burst test pressure was within 2% of that predicted by FEA.
Using TWI's pipe burst testing facility, an instrumented pipe section containing a simulated corrosion defect was tested to burst pressure (662 bar) Fig.9 to verify the accuracy of analytical failure predictions studies carried out at TWI.
Validation of the European pressure vessel code
The draft European pressure vessel code allows certain welded vessels to go into service without non-destructive testing (NDT), provided that a hydraulic test (proof test) is first carried out to demonstrate the integrity of the vessel.
TWI has carried out 24 wide plate tests as part of a European collaborative programme designed to validate the code. Wide plate specimens were made, in which the weld contained a deliberate lack of penetration flaw ( Fig.10). The plates were then subjected to a proof load of up to 1.7 times the design stress. This was followed by fatigue cycling in TWI's 100 tonnes Mayes machines to simulate the pressure cycling which would be endured in service ( Fig.11). Both carbon steel and stainless steel specimens were tested; samples were subsequently sectioned and metallographically examined to measure the extent of fatigue crack propagation. The results of these tests are combined with those carried out by other members of the consortium to provide firm recommendations for the hydraulic testing of uninspected or partially inspected pressure vessels.
Crack arrest checks for new LPG storage tank steels
A TWI Member company has been studying the possibility of using alternative materials for new liquefied propane gas (LPG) storage tanks with an operating temperature of -50°C. Previous TWI group sponsored projects have shown that the crack arrest properties of 1.5% nickel thermomechanically controlled processed (TMCP) steel and their weldments are suitable for LPG applications. However, recent trends in steel making and the high cost of nickel have resulted in development of micro-alloyed TMCP steels with nickel contents of approximately 0.5%. These steels are reported to possess very similar crack arrest properties.
To provide independent data, a test programme was carried out consisting of ten large-scale double-tension crack arrest wide plate tests ( Fig.12) on parent steels, weld metal and HAZ regions to study structurally representative crack arrest behaviour.
The results showed that there are 0.5% nickel TMCP steel grades and associated higher nickel weldments available which will arrest fast running brittle cracks under LPG storage tank operating conditions. The use of 0.5% nickel steel for LPG storage tank fabrication is therefore a possibility, but independent evaluation of properties under structurally realistic conditions is needed before any decision is made.
Validation of girth weld repair assessment procedure for offshore pipelaying
Girth welds made on laybarges are subjected to NDT, any flaws which do not meet the specified acceptance criteria are removed. Defect removal can either be by in-situ repair or by reversing the barge so that the repair or cut-out is undertaken at a location where the pipe is no longer under load. In-situ repair is desirable to improve productivity and reduce cost. However, since local repairs are conducted under loads, which may be as high as 90% of the pipe yield strength, an analysis has to be undertaken to ensure that the in-situ repair is safe.
An analysis procedure has been developed to predict safe repair groove sizes. The procedure has been validated by a series of large-scale fracture mechanics full-scale pipe bend tests on 36in (914mm) diameter API X65 linepipe (31mm wall thickness) which simulated closely the actual removal process of defects during weld repair ( Fig.13).
Fatigue testing of tubular joints
The development of the oil and gas fields in the North Sea over the past 35 years prompted very extensive research into the fatigue properties of the tubular steel jacket structures used for exploration and production. TWI was deeply involved in the resulting UKOSRP programmes, funded by the UK government with support from the European Commission, and a number of Group Sponsored Projects. Special rigs were constructed to enable large-scale tubular joints, ranging from 150mm dia x 6mm wall thickness to 1500mm dia x 65mm wall thickness, ( Fig.14), to be tested in air and sea-water, under constant and variable amplitude loading.
Such tests invariably require extensive instrumentation with strain gauges and TWI is well equipped with multi-channel dynamic loggers. Regular use was also made of AC and DC potential drop equipment for monitoring crack growth.
Large scale testing of cracked tubular joints
Flooded member detection (FMD) is one of the techniques for identifying cracking in joints offshore. This shows that a through-thickness crack has allowed water to penetrate a previously dry member. To evaluate whether the joint's performance under severe wave loading is reduced by a through-thickness crack, full size tubular cruciform specimens were fabricated at Abington and loaded in alternating tension to produce a large crack at the joint. Subsequently static tests were performed in a purpose built testing rig to find the ultimate strength of the cracked joint (see
Fig.15). Strain gauges were installed, a geometric grid was painted on the main member and measuring devices were fitted to detect ovality changes in the main tube during loading.
An understanding was gained of the relationship between crack size and the reduction in capacity of the joint to take severe wave loading.
Large-scale buckling tests to support the design for decommissioning offshore platforms
Platforms designed in the early 1960s were given a life expectancy of 30 years to coincide with gradual drying up of oil wells. There are now a number of platforms which are uneconomic to operate and which require safe and controlled abandonment. An option for decommissioning deeper water platforms is to remove all environmentally unfriendly materials and topple the structural frame. This will then end up safely on the sea bed. One technique is to cut the jacket into two halves and use two of the four legs on each half-jacket as structural hinges.
TWI has carried out an extensive testing programme within one of Europe's first major research projects into controlled demolition of disused offshore platforms. It involved the design and construction of a purpose-built computer-controlled test rig for multi-axial loading tests on tubular specimens, to induce failure.
The TWI test rig ( Fig.16) was built to buckle and bend tubular specimens. It produces forces up to 250 tonnes in axial loading, 120 tonnes in shear loading and 66 tonnes/m at the moment of in-plane bending. Tubular specimens simulating offshore jacket legs, with diameters in excess of 0.3m were fabricated and tested successfully.
Full-scale fatigue testing of risers and tendons
To investigate the fatigue behaviour of girth welded tubulars used for offshore applications such as production and export risers and tendons of tension leg platforms, fabrication and fatigue testing of many full-size girth welds has been accomplished to provide evidence for establishing acceptable new fatigue rules for one-sided girth welds to broaden their use for lower cost welds in structural tubular members and to establish the fatigue performance of flush ground girth welds.
The large resonance test rig, designed, fabricated and commissioned at TWI is capable of fatigue testing full-scale tubulars of 325mm diameter and 20mm wall thickness at a loading frequency as high as 33Hz ( Fig.17).
Conclusions
Any theoretical component design development project should be complemented by a demonstration test in which predictions can be compared with actual behaviour. In addition to extensive experience in carrying out standardised tests, many of which were based on past TWI work, TWI has been developing test techniques and test rigs for more than 50 years and has a unique combination of capabilities and staff to use specialised test techniques and test rigs to meet the need of TWI clients in a cost-effective manner.
By way of a series of examples this article has sought to demonstrate the unique mechanical testing facilities and test technique development available at TWI. Through its proven track record of rapid response to requests to carry out mechanical tests, TWI has helped many Industrial Member Companies to complete contracts on schedule.
